- Email: [email protected]
Potential impact of (CET) carbon emissions trading on China’s power sector: A perspective from different allowance allocation options
Energy 35 (2010) 3921e3931
Contents lists available at ScienceDirect
Energy journal homepage: www.elsevier.com/locate/energy
Potential impact of (CET) carbon emissions trading on China’s power sector: A perspective from different allowance allocation options Rong-Gang Cong a, b, d, Yi-Ming Wei b, c, * a
School of Business Administration, North China Electric Power University, Beijing 102206, China Center for Energy and Environmental Policy Research, Beijing Institute of Technology, Beijing, 100081, China c School of Management and Economics, Beijing Institute of Technology, Beijing, 100081, China d Institute of Policy and Management, Chinese Academy of Sciences, Beijing, 100080, China b
a r t i c l e i n f o
a b s t r a c t
Article history: Received 4 February 2010 Received in revised form 5 May 2010 Accepted 9 June 2010
In Copenhagen climate conference China government promised that China would cut down carbon intensity 40e45% from 2005 by 2020. CET (carbon emissions trading) is an effective tool to reduce emissions. But because CET is not fully implemented in China up to now, how to design it and its potential impact are unknown to us. This paper studies the potential impact of introduction of CET on China’s power sector and discusses the impact of different allocation options of allowances. Agent-based modeling is one appealing new methodology that has the potential to overcome some shortcomings of traditional methods. We establish an agent-based model, CETICEM (CET Introduced China Electricity Market), of introduction of CET to China. In CETICEM, six types of agents and two markets are modeled. We find that: (1) CET internalizes environment cost; increases the average electricity price by 12%; and transfers carbon price volatility to the electricity market, increasing electricity price volatility by 4%. (2) CET influences the relative cost of different power generation technologies through the carbon price, significantly increasing the proportion of environmentally friendly technologies; expensive solar power generation in particular develops significantly, with final proportion increasing by 14%. (3) Emissionbased allocation brings about both higher electricity and carbon prices than by output-based allocation which encourages producers to be environmentally friendly. Therefore, output-based allocation would be more conducive to reducing emissions in the Chinese power sector. Ó 2010 Published by Elsevier Ltd.
Keywords: Carbon emissions trading Emission-based allocation Output-based allocation Agent-based model
1. Introduction The EU (European Union) ETS (emission trading scheme) directive was first published in 2003, aiming at helping its members prepare for carbon emission targets agreed in the Kyoto Protocol for the 2008e2012 trading period [1]. An emission-based allocation was available in EU ETS. From January 1, 2005, if enterprises of member states are about to emit carbon dioxide, they must gain the rights to do so, which can be received from their government or bought from the carbon trading market. The main reason for the introduction of ETS is that it can reduce emissions in a cost-effective way [2]. Since January 1, 2008, emission rights can be traded among EU countries. With rapid economic development, China’s energy consumption, especially electricity consumption, has grown rapidly. More than
* Corresponding author. School of Management and Economics, Beijing Institute of Technology (BIT), 5 South Zhongguanccun Street, Haidian District, Beijing 100081, China. Tel./fax: 86 10 68911706. E-mail addresses: [email protected], [email protected] (Y.-M. Wei). 0360-5442/$ e see front matter Ó 2010 Published by Elsevier Ltd. doi:10.1016/j.energy.2010.06.013
75% of China’s electricity power is thermal power. Coal-fired electricity generation has increased rapidly over recent years. From 1991 to 2005, the proportion of coal-fired generation in thermal power was maintained at 90%e96%, meaning large amounts of carbon emissions during that period. In 2005, carbon emissions from the Chinese power sector reached 38.73% of total emissions of primary energy [3]. The Chinese power generation sector has been liberalized, allowing competition. The current electricity pricing mechanism in China is mainly based on the up-grid electricity price management approach, implemented May 1, 2005. Two-part electricity prices are applied to the generation units. The capacity price is set by the government. The electricity price is determined through market competition. The specific price is set through consultations between the power generation and power grid companies, on the premise that generation cost can be compensated. In this paper, we study the potential impact of introducing CET (carbon emissions trading) on China’s power sector. This is because, on the one hand, carbon emissions from the power sector are large; and, on the other hand, because of regional division of the power market, cross-border and inter-regional electricity transmission is
3922
R.-G. Cong, Y.-M. Wei / Energy 35 (2010) 3921e3931
relatively small, subject to transmission capacity and line loss. According to China Energy Statistical Yearbook 2008, China’s power imports in 2007 only accounted for 0.13% of the total electricity supply [4]. Introduction of CET will inevitably increase the companies’ costs and weaken their international competitiveness. As was argued above, inter-connector capacity between China and neighboring countries is relatively low and it is not expected to increase dramatically. Therefore the physical supply of electricity will remain a regional product. That means that the model here is applicable for China power sector. This paper attempts to establish an agent-based model on the introduction of CET into the Chinese power market, and attempts to answer the following questions: (1) What would be the impact of CET on the electricity price and final portfolio of power plants? (2) What would be the difference of emission-based allocation and output-based allocation? What is the micro mechanism behind it? Which allocation mechanism would be more suitable for China’s domestic carbon market? The results may shed light on the wisdom of the government’s decision and even potentially cause a change in policy. Also, some conclusions may be the general rules of carbon market, which will be a contribution to existing literatures. The article is structured as follows: firstly, we will introduce the literatures relating to the carbon market mechanism and agent-based model; secondly, a CET introduced Chinese electricity market simulation model is established; thirdly, we carry out simulations to answer the questions above; finally, a summary and some policy implications are given. 2. Literature review The literatures relating to CET mainly focus on carbon market mechanism design and its applications between and within countries. Regarding carbon market mechanism design, Bohringer and Lange discuss emission-based and output-based allocations [5]. They find that in a closed-trade system, emission-based allocation is better; however, in an open-loop system, combination of the two allocations is better. Cramton and Kerr believe that the government should auction carbon emission rights, rather than allocating them for free [6]. They argue that auction can promote technological innovation and share cost effectively. Burtraw et al. compared auction, and emission-based and output-based allocations based on the Haiku electricity market model [7,8]. They find that the social cost of auction is about half of the two free allocation types, and that emission-based allocation is more favorable to producers, while output-based allocation leads to the lowest electricity price and highest gas price. Regarding inter-state trading of emission rights, Haurie and Viguier propose a computable stochastic equilibrium model to represent possible competition between Russia and China on the international market of carbon emissions permits [9]. They analyzed the impact of this competition on the pricing of emissions permits and on the effectiveness of the Kyoto and post-Kyoto agreements, without US participation. Carlén investigates market power in intergovernmental CET based on a simulation [10]. They find that (i) the presence of a large trader is not likely to create inefficiencies, and (ii) a large trader is not likely to be able to substantially influence prices to its advantage during end-period trading. Bosello and Roson explore the distributional consequences of alternative emissions trading schemes [11]. They find that the introduction of a competitive market for emissions permits, especially when this market includes non-Annex I countries (developing countries), would dramatically lower total abatement costs but, on the other hand, would primarily benefit the richest countries.
Regarding intra-state trading of emission rights, Keats et al. believe that as a result of introducing EU ETS, the net values of both a typical PC (pulverized coal-fired) power station and a more modern gas-fired CCGT (combined cycle gas turbine) would increase [12]. They also argue that in the future, a greater proportion of allowances can and should be auctioned. Szabó et al. developed a global dynamic simulation model of the cement industry, quantifying the benefit achieved from emission trading in different markets (EU15, EU27 and Annex B) and assessing the magnitude of the potential carbon leakage effect [13]. Chappin and Dijkema present an agent-based model to elucidate the effect of CET on the decisions of power companies in an oligopolistic market [14]. They find that even after the introduction of CET, capacity expansion plans indicate a preference for coal. In power generation, the economic effect of CET is not sufficient to outweigh the economic incentives in choosing coal. ABM (Agent-based modeling) is a typical bottom-up method. Nowadays, agent-based model has received increasing attentions from many researchers in the field of energy system modeling [15]. ABM is thought as a powerful tool for studying complex adaptive systems. The applications of agent-based methods in energy market reform and energy policy simulation are mainly focused on the study of electricity market. Bunn and Oliveira used an agent-based simulation of England and Wales electricity markets to analyze the market power and the market design [16]. Based on Sandia models, Ehlen et al. constructed a multi-agent model to simulate both uniform-price and RTP (real-time price) contracts of U.S. electricity wholesalers [17]. They found that RTP contracts made power loads move from peak to off-peak hours, increased wholesalers’ profits, but also created susceptibilities to short-term market demand and price volatilities. Bunn and Oliveira developed a simulation model of technological evolution in electricity markets to analyze how the market performance depended upon the different technological types of plant owned by the generators [18]. Hamalainen et al. did a simulation of consumer coalitions in Finnish electricity retail market to analyze the response of different types of consumers to TOU (Time-Of-Use) pricing [19]. Bernal-agustin et al. presented a realistic simulator of the day-ahead electricity market in mainland Spain, which was a good decision-making tool for the electricity market participants [20]. Cong and Wei examined whether carbon allowance auction should adopt a uniform-price or discriminatoryprice format using an agent-based model [21]. They found that a discriminatory-price auction is more suitable in terms of maximizing revenue for the government, but an uniform-price auction is better in terms of fairness to bidders, especially small bidders. The crisis of California electricity market in 2000 reminds us that a new market mechanism, if it is not fully tested before practice, can often produce the unexpected impact on the entire economy. Because CET is not fully implemented in China up to now [22], in order to analyze its potential impact on China power sector under an experiment framework which can be repeated, this paper constructed an agent-based model which is named as CETICEM (CET Introduced China Electricity Market) and consider electricity price elasticity and carbon market clearing in China, which are very important in intra-state CET. We also compare emission-based allocation and output-based allocation, which is of particular significance for domestic carbon market mechanism design. 3. Methodology: CETICEM 3.1. Model settings In the model, six types of agents are modeled: a power grid company, power producers, other industries, consumers, the government and a carbon market maker. And two markets, the electricity and carbon markets are also modeled (as shown in Fig. 1).
R.-G. Cong, Y.-M. Wei / Energy 35 (2010) 3921e3931
3923
Definitions of the parameters and variables in equation (1) are shown in Tables 1 and 2. Expected profitability is based on the ratio of discounted net cash flows in each period and total investment costs. Different from previous studies, we consider the impact of carbon quota (ceri,j(t) cp(t)) and capital rate (infl). In equation (1), carbon quota rights can be seen as revenue because it is free allocated by government. At the same time, producers may prefer environmentally friendly electricity generation technology (for which mj is smaller). The construction of nuclear plants does not only depend on producers’ attitudes, but also on national policy. The construction of wind plants and hydropower stations depends on available land and water resources, respectively. The plants are graded, the ones with higher grade having a greater probability of being invested in. We assume that there are three types of producers: profithunting producers, environmentally friendly producers and neutral producers, with weights of profit-hunting and environmental friendliness as follows: (0.7, 0), (0.3, 0.4) and (0.5, 0.2), respectively. For example, the weight on the expected profitability for profithunting producers is 0.7, while it is 0.3 for environmentally friendly
3.2. The action rules setting 3.2.1. Power producers Power producers aim to maximize their profits. The main actions of the power producers can be seen in Fig. 2. At the beginning of each period, power producers obtain carbon quotas and determine their electricity supplies. Based on the carbon needed for electricity generation, producers decide to buy or sell carbon quotas. At the end of each period, based on the supply and demand of the market and of the operations themselves, the producers make decisions to establish or close power plants. Which type of plant will be invested in is based on multicriteria. The size of plant will be decided according to the criteria for meeting the electricity demand. The criteria include economic, environmental friendliness, nuclear fear, and the constraints of land and water resources. The economic criterion indicates the expected profitability of a power plant. The expected profitability of producer i’s plant j, epi,j, can be calculated in equation (1):
ceri;j ðtÞ cpðtÞ þ capaj pðtÞ capaj nj fpj ðtÞ capaj mj cpðtÞ ocj 1, ltj 1 A incoj epi;j ðtÞ ¼ @ 1 ð 1 þ infl þ subj incoj 1 1 1 þ infl 0
Fuel market
Other industries Carbon quotas Set quotas
Carbon market
Fuel
Cost
Price
Current installation: information Coal-fired power plants Carbon quotas
Supervise
Supply
Hydropower station…
Subsidy/Tax Government
Power producers
Set quotas
Supply
Adjust
Electricity market
Available installation: Invest
Supply Demand Price Information
Fuel price information
Price information
Supervise
Foreign power producers
Wind power plant
decision Demand information Price information
Coal-fired power plants Gas-fired power plants Solar power plant Wind power plant Hydropower Station
Power grid companies
Nuclear Power Plant Supply
Demand Consumers
Scenario generator Physical flow
Information flow Fig. 1. System structure.
Exogenous agents
(1)
3924
R.-G. Cong, Y.-M. Wei / Energy 35 (2010) 3921e3931
Obtain carbon quotas
Determine the electricity supply which is shown in the section “electricity market”
Buy fuels
Calculate carbon needed for generation
No
Carbon needed
Buy quotas
>Available quota
Yes
Is there a shortage of
No
electricity supply?
Sell quotas
Yes Multi-criterion decision
No
Is the plant at end of its life
or
making
a
continuous loss?
Yes Close the plant
Fig. 2. Flow chart of power producers’ actions.
producers. The weight on the expected profitability for neutral producers is 0.5, which means their relatively neutral attitudes to the economic criterion. In reality, profit-hunting producers may refer to private power companies whose objective is only profit maximization. Environmentally friendly producers may refer to public power companies. Not only profit maximization but also environmental protection is their objectives. Neutral producers fall in between two types of producers above. While investing, producers will firstly decide their type of producer according to the probability vector Pi ¼ {pp,pe,pn}, where pp,pe,pn represent the probabilities of turning to profit-hunting producers, environmentally friendly producers and neutral producers. They will then adjust their vectors at the end of each period based on the rules described in Table 3. For example, when
producer’s type is profit-hunting, if it can make profits, the probability of turning to profit-hunting is increased by D, the other two probabilities are decreased by D/2. The rules of probability adjusting are inspired by reinforcementlearning theory where agents take actions to maximize their longterm utilities. So the action which can get larger utility is more likely adopted in the future. Through the reinforcement-learning mechanism mentioned above, we can study whether different market design will shift consumers’ preferences. 3.2.2. Government The government’s main aim is to reduce emissions and stabilize the electricity price. Its actions include subsidizing environmentally friendly technology, laying the carbon allocation plan, setting the
R.-G. Cong, Y.-M. Wei / Energy 35 (2010) 3921e3931 Table 1 Description of parameters.
Table 3 The rules of probability adjusting.
Parameter Description
Parameter Description
a
b
g q capaj Infl mj
nj subj
GDP elasticity of electricity demand The decrease in the rate of the quota Proportional coefficient Capacity of electricity plant j Capital interest rate Carbon needed for plant j’s unit of electricity Fuel needed for plant j’s unit of electricity Subsidy for plant j, which is environmentally friendly
l ave_mar incoj ltj mari
ocj
Price elasticity of electricity demand Proportional coefficient The average profit rate of the electricity industry Investment cost for plant j Plant j’s life The ratio of producer i’s profit rate and the electricity industry’s profit rate Operation cost for plant j in one period
Producer’s type when investing
Whether the plant can make profits?
Probability adjusting
Profit-hunting producers Environmentally friendly producers Neutral producers
Yes No Yes No Yes No
P)fpp þ D; pe D2 ; pn D2 g P)fpp D; pe þ D2 ; pn þ D2 g P)fpp D2 ; pe þ D; pn D2 g P)fpp þ D2 ; pe D; pn þ D2 g P)fpp D2 ; pe D2 ; pn þ Dg P)fpr þ D2 ; pe þ D2 ; pn Dg
su ðt 1Þ ceri ðtÞ ¼ P i to capðtÞ g sui ðt 1Þ
sui(t 1) is producer i’s electricity supply in period t 1. In this case, ceri,j(t) in (1) is defined as equation (5):
sui;j ðt 1Þ ceri;j ðtÞ ¼ P P to capðtÞ g sui;j ðt 1Þ
national nuclear policy and controlling the electricity price. Carbon allocation in our model focuses on two of these actions: (I) Emission-based allocation: The previous year’s emission proportion is the standard used for setting emission quotas in the following year, which is the current quota allocation mechanism in the EU ETS. Producer i’s emission quota, ceri(t), can be calculated in equation (2):
emi ðt 1Þ ceri ðtÞ ¼ P i to capðtÞ g emii ðt 1Þ
(5)
j
And the relationship between ceri(t) and ceri,j(t) is shown in equation (4):
ceri ðtÞ ¼
X
ceri;j ðtÞ
j
Although in the first phase of EU ETS, other allocation modes, such as auction, have been in practice. But there were only three countries (Ireland, Hungary and Lithuania) choosing auction. And the percentages of carbon allowances auctioned are relatively few. Therefore, we choose two free grandfathering allocations which drew the most attentions and try to compare them.
(2)
i
where emii(t 1) is producer i’s carbon emissions in period t 1; to_cap(t) is total carbon quota in period t; and g is the decrease in the rate of the quota. ceri,j(t) in (1) is defined as equation (3):
mj sui;j ðt 1Þ ceri;j ðtÞ ¼ P P to capðtÞ g mj sui;j ðt 1Þ
(4)
i
i
i
3925
(3)
j
where sui,j(t 1) is electricity supply of producer i’s plant j in period t 1. (II) Output-based allocation: The supply proportion of the previous year’s electricity is the standard used for setting emission quotas for the following year, as shown in equation (4):
3.2.3. Industry The industry’s aim is to sell products to maximize profit. We use an agent to represent industry and assume its demand is determined by an exogenous scenario. Because power sector is our focus, industry agent’s carbon demand is set exogenously here. We will expand our object of study in the future. 3.2.4. Power grid company The action of the power grid company is setting the transmission and distribution tariff. In 2007, the average transmission and distribution tariff was 160.12 yuan/KKWH, which accounted for 31.49% of the electricity price [23]. Therefore, in the model retail electricity price equals to 131.49% of average electricity price. 3.2.5. Consumers According to different sale prices, we divide consumers into two categories: small and large electricity consumers. Small consumers need to bear the electricity price, and large consumers need to
Table 2 Description of variables. Variables
Description
Variables
Description
ave_ep(t) bid_ave(t) car_s(t) ceri,j(t) D(t) epi,j(t) GDP(t) pool(t) sui,j(t) total_cap(t) to_cap(t)
The average electricity price The average market price Total carbon supply in period t Carbon quota of producer i’s plant j Electricity demand in period t The electricity price for producer i’s plant j GDP in period t The cumulative imbalance in period t Electricity supply of producer i’s plant j in period t Total available quotas in period t Total carbon quota
bidi,j(t) car_d(t) ceri(t) cp(t) emii(t) fpj(t) P(t) si,j(t) sui(t) total_emission(t) w_sij(t)
The bid price of plant j owned by producer i in period t Total carbon demand in period t Producer i’s emission quota Carbon price in period t Producer i’s carbon emissions in period t Fuel price for plant j Retail electricity price in period t The actual supply of plant j owned by producer i in period t Producer i’s electricity supply in period t Emissions in period t The electricity supply of producer i’s plant j in period t
3926
R.-G. Cong, Y.-M. Wei / Energy 35 (2010) 3921e3931
Fig. 3. Electricity prices corresponding to different mechanisms (GDP growth, 3%; interest rate, 3%; reduction of quota, 2%). Note: we define GDP growth is 3%, because it will show the convergence trend in the long term. We also do the sensitivity analysis for different GDP and get the similar results.
cover the electricity price and capacity price. We use two agents to represent consumers. China’s electricity demand is mainly driven by gross domestic product (GDP). Meanwhile, as a commodity, its demand is certainly affected by its price. Therefore, we assume electricity demand is decided by GDP growth and its price changes accordingly. The consumers’ decisions are shown in equation (6):
3.2.6. Carbon market maker The primary function of the carbon market maker is to provide market liquidity and balance the supply and demand of the carbon market. The cumulative imbalance in period t, pool(t), can be calculated as in equation (7):
GDPðtÞ D t ¼ Dðt 1Þ 1 þ 1 a 1 GDPðt 1Þ PðtÞ 1 b þ Pðt 1Þ
where car_d(t) refers to total carbon demand in period t; car_s(t) is total carbon supply in period t; and pool(t) > 0 represents cumulative CO2 demand is larger than CO2 supply, and vice versa.
(6)
where GDP(t) refers to GDP in period t; P(t) represents retail electricity price in period t; a is the GDP elasticity of electricity demand; and b is the retail price elasticity of electricity demand.
Table 4 Comparison of electricity prices under different mechanisms. Electricity price
CET introduced
No CET introduced
Ratio
Electricity price at the end of 100 periods (yuan/KKWH) Maximum electricity price (yuan/KKWH) Minimum electricity price (yuan/KKWH) Average electricity price (yuan/KKWH) Standard deviation (yuan/KKWH) Standard deviation coefficient
1162.6
947.26
1.23
1170.38 486 876.44 176.10 0.20
965.61 486 780.72 150.13 0.19
1.12 1.3 1.04
poolðtÞ ¼ poolðt 1Þ þ car dðtÞ car ðtÞ
(7)
3.3. The market rules 3.3.1. Electricity market A two-part electricity price system operates in China, comprising the electricity price and capacity price. The capacity price is based on the average investment cost of generating units, which aims to compensate for investment cost and finance charges. The electricity price is determined by the market. This paper sets the rules as follows: in period t, firstly, each plant decides the electricity supply, w_si,j(t). If the expected return of generation is less than selling the carbon quota, the electricity supply is zero; otherwise, the supply equals its capacity. The average market price, bid_ave(t), can be calculated in equation (8):
Pn bid aveðtÞ ¼
i¼1
Pm
Pn
w si;j ðtÞ bidi;j ðtÞ Pm j ¼ 1 w si;j ðtÞ
j¼1
i¼1
Fig. 4. Electricity price at the end of 100 periods corresponding to different GDP growth rates.
(8)
R.-G. Cong, Y.-M. Wei / Energy 35 (2010) 3921e3931
3927
Fig. 5. Power source structure when CET is introduced.
As shown in equation (8), the average market price is weighed to the capacity of the bids. The bid price of plant j owned by producer i is composed of cost and profit, shown in equation (9):
bidi;j ðtÞ ¼ fpj ðtÞ capaj ðtÞ þ ocj ðtÞ þ capaj ðtÞ mj cpðtÞ ð1 þ mari ave marÞ=capaj ðtÞ (9) where ave_mar denotes the average profit rate of the electricity industry, which is set exogenously; and mari is the ratio of producer i’s profit rate and the electricity industry’s profit rate. Based on the argument that the bidding price determines the amount of capacity sold, the actual supply of plant j owned by producer i is shown in equation (10):
si;j ðtÞ ¼ w si;j ðtÞ Pn
i¼1
DðtÞ Pm j¼1
w si;j ðtÞ
bid avei ðtÞ bidi;j ðtÞ
(10)
In period t, the average electricity price, ave_ep(t), is shown in equation (11):
ave epðtÞ ¼ bid aveðtÞ e
Pn
q to DðtÞ
i¼1
Pm j¼1
w si;j ðtÞ
where q is a proportional coefficient, which reflects price fluctuations when there is an imbalance between supply and demand. As mentioned above, the bids of power plants determine the frequency of their usage. In reality, the actual price received for the electricity produced by a specific power plant is related to the height of its bid price. The electricity price for each plant is shown in equation (12):
epi;j ðtÞ ¼ ave epðtÞ
bidi;j ðtÞ bid aveðtÞ
Some validations of equations above can be found in Chappin’s research [24]. 3.3.2. Carbon market We assume that the carbon price is affected by carbon supply and demand. The current carbon price is affected by the total available quotas and emissions of the last period (current emission is not available). In period t, carbon price, cp(t), is as shown in equation (13):
cpðtÞ ¼ cpð0Þ eðlðtotal (11)
(12)
emissionðt1Þtotal capðtÞþpoolðtÞÞÞ
ð1 þ et Þ; et wN 0; 0:012
Gas
Coal
Wind
Solar Nuclear Hydro
Fig. 6. Power source structure when CET is not introduced.
(13)
3928
R.-G. Cong, Y.-M. Wei / Energy 35 (2010) 3921e3931
Table 5 Comparison of power source structures under different schemes. CET introduced The beginning period of large-scale gas 2 power generation (Proportion > 8%) Final portfolio (%) Hydropower 6.83 Solar power 24.27 Wind power 0.31 Solar power 25.25 Coal-fired 17.79 power Gas-fired power 25.53 The absolute amount Hydropower 441477.8 (MW) Wind power 20225
No CET introduced 34 11.13 16.51 0.50 11.10 36.13 24.64 453848.2 20205.9
where l is a proportional coefficient. In addition, a random disturbance, et, is also added to the model to explain other random factors and cp(0) is the initial value for the carbon price, which is set to 100 according to the EU ETS and exchange rate. 3.4. Model implementation The software used here is NetLogo which is a cross-platform multiagent programmable modeling environment. We created 90 agents as electricity power producers, each producer having a number of plants. Relevant parameters’ value of power plants are shown in the appendix Table 1. The agents evolve over time by action and interaction. The total simulation length is 100 periods. The model procedure is repeated until the end of the simulated period is reached.
grid price for wind power is 617.58 yuan/KKWH; for nuclear power is 436.23 yuan/KKWH; for thermal power is 346.33 yuan/KKWH; for hydropower is 244.04 yuan/KKWH. We can see from Fig. 3 that compared to no CET, introducing CET has the following effects: increase in the level and fluctuation of the electricity price in most cases; at the end of 100 periods (here one period means one year), the electricity price is 1.23 times higher; and fluctuation of the electricity price is 1.04 times higher (see Table 4). Therefore, introducing CET internalizes costs and transfers the volatility of the carbon price to the electricity market. Next, we study the impact of CET on the electricity price under different GDP growth rates. We do a simulation of the electricity price at the end of 100 periods corresponding to different GDP growth rates. The results are shown in Fig. 4. We can see from Fig. 4 that if no CET is introduced, when the GDP growth rate is low, the electricity price growth rate is high; and when the GDP growth rate is high, the electricity price growth rate shows a decreasing trend. On one hand, economic development encourages electricity consumption and drives up the electricity price. On the other hand, increase in the electricity price suppresses consumption. Therefore, economic development has both positive and negative impact on electricity consumption, which in return reflects the trend of decrease in electricity price growth. However, electricity price growth shows an increasing trend when CET is introduced. This is because, in addition to the above impact, economic development also demands carbon, which pushes up the carbon price. The carbon price increase is reflected in the electricity price, pushing up the electricity price. Therefore, to address China’s growing economy, the Chinese government should establish regulatory measures if CET will be introduced.
4. Experiment design and analysis 4.2. Impact of CET on power source structure The main assumptions underlying the model are an electricity market with perfect competition, a static number of electricity producers and a limited available power plant types. Main input data are listed in appendix Table 2. 4.1. Impact of CET on electricity price First of all, we explore the impact of introducing CET on the electricity price in China. The emission-based allocation, which was used in EU ETS, is chose here as the baseline scenario. We obtain different electricity prices corresponding to whether CET is introduced or not. With due consideration of availability of data, 2007 is chose as the base year. Data from China state electricity regulatory commission are used to calibrate the model. In 2007, average on-
Next, we consider the impact of CET on China’s power source structure. The emission-based allocation is also chose here. We can see from Figs. 5 and 6 that introducing CET brings forward large-scale natural gas power generation. At the same time, we can see from the final power source structure (Table 5) that there is an increase in the proportion of environmentally friendly power generation technology (nuclear power, solar power and gas-fired power). Among them, solar power, which provides environmental protection but is expensive, develops a lot after introducing CET. Due to land and water constraints, the proportions of wind power and hydropower, respectively, do not increase after CET is introduced, although, the absolute amount of wind power increases slightly.
Fig. 7. Electricity price and carbon price of two carbon quota allocations.
R.-G. Cong, Y.-M. Wei / Energy 35 (2010) 3921e3931
3929
Table 6 Group statistics of economic producers. Type of producer
Mean
Std. deviation
Minimum
Maximum
Emission-based Output-based
0.394 0.328
0.053 0.0534
0.3 0.2
0.6 0.5
Table 7 Independent sample t-test for proportion of economic producers.
Projected proportion of profit-hunting Producers
Equal variances assumed Equal variances not assumed
Levene’s test for equality of variances
t-test for equality of means
F
df
Sig.
t
0.369 0.544 8.845 198
Sig. Mean (2difference tailed)
Fig. 8. Impact of introducing CET on final producers’ proportions.
0.000 0.0666
8.845 197.989 0.000 0.0666
Table 8 Group statistics of environmentally friendly producers. Type of producer
Mean
Std. deviation
Minimum
Maximum
Emission-based Output-based
0.266 0.327
0.048 0.051
0.14 0.21
0.44 0.48
4.3. Comparison of emission-based allocation and output-based allocation As can be seen from Fig. 7, the evolution of the electricity price and carbon price of the two allocation mechanisms is similar. However, the electricity price of output-based allocation is lower because of the lower carbon price. We would like to know what the micro mechanism for this macro phenomenon is and the difference of producers under both market mechanisms. We analyze the final producers’ proportions according to the type of allocation, i.e., emission-based and output-based allocations (initial proportions: economic producers, 0.3; environmentally friendly producers, 0.3; neutral producers, 0.4). The results are as follows: As can be seen from Table 6, when emission-based allocation is introduced, the proportion of economic producers (0.394) is larger than the proportion when output-based allocation is introduced (0.328). As seen from Table 7, the probability of F is 0.544 for projected proportion of profit-hunting producers, which is insignificant at the 5% level. Thus, we believe that equal variances assumed under two allocations are not rejected. Next, let us look at the results corresponding to assumptions of equal variances. The value of t is 8.845, with probability is 0.000. The result rejects the null hypothesis and shows that there is a significant difference between the proportions of economic producers under the two allocation mechanisms. Table 9 Group statistics of neutral producers. Type of producer
Mean
Std. deviation
Minimum
Maximum
Emission-based Output-based
0.339 0.345
0.051 0.0538
0.19 0.22
0.456 0.49
We do a similar analysis on neutral and environmentally friendly producers (intermediate results are shown in Tables 8 and 9). The robust test is also done for different initial proportion settings. The final results are shown in Fig. 8. As seen in Fig. 8, when emission-based allocation is introduced, the proportion of economic producers (0.394) is higher compared to output-based allocation (0.328). In other words, emission-based allocation causes producers to turn to less environmentally friendly technologies. In essence, emission-based allocation allocates more carbon quotas to producers who emit more. This is an incentive for emissions. From a broader sense, emission-based allocation is designed for compensating bodies which are affected by emission control [25e27]. Output-based allocation encourages producers to be environment-friendly. Therefore, total emissions are lower, resulting in a lower carbon price. In this perspective, in the Chinese domestic carbon market, for the same sector, output-based allocation is more efficient for reducing emissions. The results here may be different from the general opinion which was established in standard market theory that the method of allocation of a given number of emission certificates should not affect the equilibrium price. Because the producers in our model can shift their investment preferences in the two allowance allocations, output-based allocation makes producers tend to be environment-friendly, which will decrease total carbon demand and carbon price. 5. Conclusions and future work In this paper, we studied the impact of introduction of CET on China’s power sector using an agent-based model. The main findings are as follows: (1) CET internalizes the external environmental cost, which increases the average electricity price (12%). At the same time, the volatility of the carbon market is also transferred to the electricity market (fluctuations excluding the impact of mean increases of 4%). When there is no introduction of CET, the electricity price shows a declining trend as the economy develops. However, introduction of CET significantly increases the electricity price. Therefore, to address China’s high-speed development, the Chinese government should establish specific regulatory measures after introducing CET. (2) CET influences the relative costs of different power generation technologies through the carbon price, which would have a significant impact on the power source structure.
3930
R.-G. Cong, Y.-M. Wei / Energy 35 (2010) 3921e3931
Introduction of CET would cause large-scale natural gas generation. The final proportion of environmentally friendly power generation technologies such as nuclear power and natural gas power would increase. Environmentally friendly solar power would develop significantly after CET is introduced (final proportion, 14% increase). In contrast, the final proportion of coal-fired power, whose emissions are high, would decrease significantly by 18%. (3) Emission-based allocation produces both a higher electricity price and higher carbon price, compared to output-based allocation. This is because under the latter allocation, producers would tend to be more environmentally friendly. Compared to emission-based allocation, output-based allocation would be more conducive to environmental protection. Therefore, output-based allocation should be considered in the design of the Chinese domestic CET market. As we know, this paper is the first one which studied the potential impact of CET on China power sector in a computable framework. It will provide China government and related decisionmakers a quantity tool for designing carbon market. This paper also firstly explains which free allocation is better and the micro mechanism behind it, which is a necessary supplement to existing literatures. This paper does not endogenously model the Chinese fuel market. We also do not consider entry and quit mechanisms for producers. Demand elasticity of electricity is constant. These are left to address in future research. Acknowledgements The work in this paper is part of “National carbon market mechanisms design” which aims to promote China’s reduction. The authors gratefully acknowledge the financial support from the National Natural Science Foundation of China under grant Nos.70733005, 70701032. We also would like to thank Professor Henrik Lund and the anonymous referees for their helpful suggestions and corrections on the earlier draft of our paper according to which we improved the content. Appendix Table 1. Parameters’ values of power plants.
References [1] EC. Directive 2003/87/EC Establishing a scheme for greenhouse emission allowance trading within the community and amending council directive 96/ 61/EC, Report; 2003 See also, http://europa.eu/legislation_summaries/energy/ european_energy_policy/l28012_en.htm#AMENDINGACT. [2] Svendsen GT. The idea of global CO2 trade. European Environment 1999;9:232e7. [3] Wei YM, Liu LC, Fan Y, Wu G. China energy report: CO2 emissions research. Beijing: Science Press. 2008 See also, http://www.sciencep.com/s_single.php?id¼14944. [4] NBSdNational Bureau of Statistics of China. China energy statistical yearbook 1997e2008. Beijing: China Statistical Press; 2008. [5] Bohringer C, Lange A. On the design of optimal grandfathering schemes for emission allowances. European Economic Review 2005;49:2041e55. [6] Cramton P, Kerr S. Tradeable carbon permit auctions: how and why to auction not grandfather. Energy Policy 2002;30:333e45. [7] Burtraw D, Palmer K, Bharvirkar R, Paul A. The effect of allowance allocation on the cost of carbon emission trading. RFF Discussion Paper; 2001. [8] Paul A, Burtraw D. The RFF haiku electricity market model. RFF Discussion Paper; 2002. [9] Haurie A, Viguier L. A stochastic dynamic game of carbon emissions trading. Environmental Modeling and Assessment 2003;8:239e48. [10] Carlén B. Market power in international carbon emissions trading: a laboratory test. The Energy Journal 2003;24:1e26. [11] Bosello F, Roson R. Carbon emissions trading and equity in international agreements. Environmental Modeling and Assessment 2002;7:29e37. [12] Keats K, Neuhoff K, Martinez KK. Allocation of carbon emission certificates in the power sector: how generators profit from grandfathered rights. Climate Policy 2005;5:61e78. [13] Szabó L, Hidalgo I, Ciscar JC, Soria A. CO2 emission trading within the European Union and Annex B countries: the cement industry case. Energy Policy 2006;34:72e87. [14] Chappin EJL, Dijkema GPJ. On the impact of CO2 emission-trading on power generation emissions. Technological Forecasting and Social Change 2009;76:358e70. [15] Ma T, Nakamori Y. Modeling technological change in energy systemsefrom optimization to agent-based modeling. Energy 2009;34:873e9. [16] Bunn DW, Oliveira FS. Evaluating individual market power in electricity markets via agent-based simulation. Annals of Operations Research 2003;121(1):57e77. [17] Ehlen MA, Scholand AJ, Stamber KL. The effects of residential real-time pricing contracts on transco loads, pricing, and profitability: simulations using the NABLE(TM) agent-based model. Energy Economics 2007;29(2):211e27. [18] Bunn DW, Oliveira FS. Agent-based analysis of technological diversification and specialization in electricity markets. European Journal of Operational Research 2007;181(3):1265e78. [19] Hamalainen RP, Mantysaari J, Ruusunen J, Pierre-Olivier P. Cooperative consumers in a deregulated electricity market e dynamic consumption strategies and price coordination. Energy 2000;25(9):857e75. [20] Bernal-Agustin JL, Contreras J, Martin-Flores R, Conejo AJ. Realistic electricity market simulator for energy and economic studies. Electric Power Systems Research 2007;77:46e54. [21] Cong RG, Wei YM. Auction design for the allocation of carbon emission allowances: uniform or discriminatory price? International Journal of Energy and Environment 2010;1(3):533e46. [22] Raufer R, Li S. Emissions trading in China: a conceptual ‘leapfrog’ approach? Energy 2009;34(7):904e12.
Type
Life (ltj)
Installed capacity (capaj) (MW)
Unit cost of power generation (fpj(t)nj) (yuan/KKWH)
Construction cost (incoj) (yuan/KKWH)
Carbon emission (mj) (ton CO2/KKWH)
Hydrostation Nuclear plant Wind plant Solar plant Coal-fired plant Gas-fired plant
30 40 20 20 20 20
600 300 10 20 300 120
240 440 620 1900 350 800
6000 12,000 8000 10,000 4000 4000
0 0 0 0 1.3 0.7
Source: [28,29]. Some data are from the historical data of plants that have been established. Others are authors’ calculations according to actual survey data from Chinese power plants.
Table 2. Main input data. Parameter
Value
Parameter
Value
a l q
0.8 0.005 0.001 0.03 0 100
b g
0.1 0.95 0.1 1 0
Infl ocj cp(0)
ave_mar mari subj
R.-G. Cong, Y.-M. Wei / Energy 35 (2010) 3921e3931 [23] SERCdState Electricity Regulatory Commission. Monitoring report on the implementation of tariff 2008. Internal Report; 2008. See also, http://www. serc.gov.cn/zwgk/jggg/200809/W020080912334874610579.doc. [24] Chappin EJL. Carbon dioxide emission trade impact on power generation portfolio, agent-based modelling to elucidate influences of emission trading on investments in Dutch electricity generation. Master Thesis, Delft University of Technology, Delft; 2006. [25] Tietenberg T. The tradable permits approach to protecting the commons: lessons for climate change. Oxford Review of Economic Policy 2003;19:400e19. [26] Harrison D, Radov D. Evaluation of alternative initial allocation mechanisms in a European Union greenhouse gas emissions allowance trading scheme. Prepared for DG Environment, European Commission; 2002. [27] Ahman M, Burtraw D, Kruger J, Zetterberg L. A ten year rule to guide the allocation of EU emission allowances. Energy Policy 2006;35:1718e30.
3931
[28] Peng C. China hydropower project in the 21st century. Beijing: China Water Power Press; 2006. [29] SERCdState Electricity Regulatory Commission. Electricity regulatory annual report. Internal Report; 2007 See also, http://www.serc.gov.cn/zwgk/jggg/ 200808/t20080801_9841.htm.
Mr. Rong-Gang Cong is a Ph.D. candidate in Management Science at the Institute of Policy and Management, Chinese Academy of Sciences, China.
Dr. Yi-Ming Wei is a Professor at the Center for Energy and Environmental Policy Research, Beijing Institute of Technology (BIT), and is the Dean of School of Management and Economics, BIT. He was a visiting scholar at Harvard University, USA.
Contents lists available at ScienceDirect
Energy journal homepage: www.elsevier.com/locate/energy
Potential impact of (CET) carbon emissions trading on China’s power sector: A perspective from different allowance allocation options Rong-Gang Cong a, b, d, Yi-Ming Wei b, c, * a
School of Business Administration, North China Electric Power University, Beijing 102206, China Center for Energy and Environmental Policy Research, Beijing Institute of Technology, Beijing, 100081, China c School of Management and Economics, Beijing Institute of Technology, Beijing, 100081, China d Institute of Policy and Management, Chinese Academy of Sciences, Beijing, 100080, China b
a r t i c l e i n f o
a b s t r a c t
Article history: Received 4 February 2010 Received in revised form 5 May 2010 Accepted 9 June 2010
In Copenhagen climate conference China government promised that China would cut down carbon intensity 40e45% from 2005 by 2020. CET (carbon emissions trading) is an effective tool to reduce emissions. But because CET is not fully implemented in China up to now, how to design it and its potential impact are unknown to us. This paper studies the potential impact of introduction of CET on China’s power sector and discusses the impact of different allocation options of allowances. Agent-based modeling is one appealing new methodology that has the potential to overcome some shortcomings of traditional methods. We establish an agent-based model, CETICEM (CET Introduced China Electricity Market), of introduction of CET to China. In CETICEM, six types of agents and two markets are modeled. We find that: (1) CET internalizes environment cost; increases the average electricity price by 12%; and transfers carbon price volatility to the electricity market, increasing electricity price volatility by 4%. (2) CET influences the relative cost of different power generation technologies through the carbon price, significantly increasing the proportion of environmentally friendly technologies; expensive solar power generation in particular develops significantly, with final proportion increasing by 14%. (3) Emissionbased allocation brings about both higher electricity and carbon prices than by output-based allocation which encourages producers to be environmentally friendly. Therefore, output-based allocation would be more conducive to reducing emissions in the Chinese power sector. Ó 2010 Published by Elsevier Ltd.
Keywords: Carbon emissions trading Emission-based allocation Output-based allocation Agent-based model
1. Introduction The EU (European Union) ETS (emission trading scheme) directive was first published in 2003, aiming at helping its members prepare for carbon emission targets agreed in the Kyoto Protocol for the 2008e2012 trading period [1]. An emission-based allocation was available in EU ETS. From January 1, 2005, if enterprises of member states are about to emit carbon dioxide, they must gain the rights to do so, which can be received from their government or bought from the carbon trading market. The main reason for the introduction of ETS is that it can reduce emissions in a cost-effective way [2]. Since January 1, 2008, emission rights can be traded among EU countries. With rapid economic development, China’s energy consumption, especially electricity consumption, has grown rapidly. More than
* Corresponding author. School of Management and Economics, Beijing Institute of Technology (BIT), 5 South Zhongguanccun Street, Haidian District, Beijing 100081, China. Tel./fax: 86 10 68911706. E-mail addresses: [email protected], [email protected] (Y.-M. Wei). 0360-5442/$ e see front matter Ó 2010 Published by Elsevier Ltd. doi:10.1016/j.energy.2010.06.013
75% of China’s electricity power is thermal power. Coal-fired electricity generation has increased rapidly over recent years. From 1991 to 2005, the proportion of coal-fired generation in thermal power was maintained at 90%e96%, meaning large amounts of carbon emissions during that period. In 2005, carbon emissions from the Chinese power sector reached 38.73% of total emissions of primary energy [3]. The Chinese power generation sector has been liberalized, allowing competition. The current electricity pricing mechanism in China is mainly based on the up-grid electricity price management approach, implemented May 1, 2005. Two-part electricity prices are applied to the generation units. The capacity price is set by the government. The electricity price is determined through market competition. The specific price is set through consultations between the power generation and power grid companies, on the premise that generation cost can be compensated. In this paper, we study the potential impact of introducing CET (carbon emissions trading) on China’s power sector. This is because, on the one hand, carbon emissions from the power sector are large; and, on the other hand, because of regional division of the power market, cross-border and inter-regional electricity transmission is
3922
R.-G. Cong, Y.-M. Wei / Energy 35 (2010) 3921e3931
relatively small, subject to transmission capacity and line loss. According to China Energy Statistical Yearbook 2008, China’s power imports in 2007 only accounted for 0.13% of the total electricity supply [4]. Introduction of CET will inevitably increase the companies’ costs and weaken their international competitiveness. As was argued above, inter-connector capacity between China and neighboring countries is relatively low and it is not expected to increase dramatically. Therefore the physical supply of electricity will remain a regional product. That means that the model here is applicable for China power sector. This paper attempts to establish an agent-based model on the introduction of CET into the Chinese power market, and attempts to answer the following questions: (1) What would be the impact of CET on the electricity price and final portfolio of power plants? (2) What would be the difference of emission-based allocation and output-based allocation? What is the micro mechanism behind it? Which allocation mechanism would be more suitable for China’s domestic carbon market? The results may shed light on the wisdom of the government’s decision and even potentially cause a change in policy. Also, some conclusions may be the general rules of carbon market, which will be a contribution to existing literatures. The article is structured as follows: firstly, we will introduce the literatures relating to the carbon market mechanism and agent-based model; secondly, a CET introduced Chinese electricity market simulation model is established; thirdly, we carry out simulations to answer the questions above; finally, a summary and some policy implications are given. 2. Literature review The literatures relating to CET mainly focus on carbon market mechanism design and its applications between and within countries. Regarding carbon market mechanism design, Bohringer and Lange discuss emission-based and output-based allocations [5]. They find that in a closed-trade system, emission-based allocation is better; however, in an open-loop system, combination of the two allocations is better. Cramton and Kerr believe that the government should auction carbon emission rights, rather than allocating them for free [6]. They argue that auction can promote technological innovation and share cost effectively. Burtraw et al. compared auction, and emission-based and output-based allocations based on the Haiku electricity market model [7,8]. They find that the social cost of auction is about half of the two free allocation types, and that emission-based allocation is more favorable to producers, while output-based allocation leads to the lowest electricity price and highest gas price. Regarding inter-state trading of emission rights, Haurie and Viguier propose a computable stochastic equilibrium model to represent possible competition between Russia and China on the international market of carbon emissions permits [9]. They analyzed the impact of this competition on the pricing of emissions permits and on the effectiveness of the Kyoto and post-Kyoto agreements, without US participation. Carlén investigates market power in intergovernmental CET based on a simulation [10]. They find that (i) the presence of a large trader is not likely to create inefficiencies, and (ii) a large trader is not likely to be able to substantially influence prices to its advantage during end-period trading. Bosello and Roson explore the distributional consequences of alternative emissions trading schemes [11]. They find that the introduction of a competitive market for emissions permits, especially when this market includes non-Annex I countries (developing countries), would dramatically lower total abatement costs but, on the other hand, would primarily benefit the richest countries.
Regarding intra-state trading of emission rights, Keats et al. believe that as a result of introducing EU ETS, the net values of both a typical PC (pulverized coal-fired) power station and a more modern gas-fired CCGT (combined cycle gas turbine) would increase [12]. They also argue that in the future, a greater proportion of allowances can and should be auctioned. Szabó et al. developed a global dynamic simulation model of the cement industry, quantifying the benefit achieved from emission trading in different markets (EU15, EU27 and Annex B) and assessing the magnitude of the potential carbon leakage effect [13]. Chappin and Dijkema present an agent-based model to elucidate the effect of CET on the decisions of power companies in an oligopolistic market [14]. They find that even after the introduction of CET, capacity expansion plans indicate a preference for coal. In power generation, the economic effect of CET is not sufficient to outweigh the economic incentives in choosing coal. ABM (Agent-based modeling) is a typical bottom-up method. Nowadays, agent-based model has received increasing attentions from many researchers in the field of energy system modeling [15]. ABM is thought as a powerful tool for studying complex adaptive systems. The applications of agent-based methods in energy market reform and energy policy simulation are mainly focused on the study of electricity market. Bunn and Oliveira used an agent-based simulation of England and Wales electricity markets to analyze the market power and the market design [16]. Based on Sandia models, Ehlen et al. constructed a multi-agent model to simulate both uniform-price and RTP (real-time price) contracts of U.S. electricity wholesalers [17]. They found that RTP contracts made power loads move from peak to off-peak hours, increased wholesalers’ profits, but also created susceptibilities to short-term market demand and price volatilities. Bunn and Oliveira developed a simulation model of technological evolution in electricity markets to analyze how the market performance depended upon the different technological types of plant owned by the generators [18]. Hamalainen et al. did a simulation of consumer coalitions in Finnish electricity retail market to analyze the response of different types of consumers to TOU (Time-Of-Use) pricing [19]. Bernal-agustin et al. presented a realistic simulator of the day-ahead electricity market in mainland Spain, which was a good decision-making tool for the electricity market participants [20]. Cong and Wei examined whether carbon allowance auction should adopt a uniform-price or discriminatoryprice format using an agent-based model [21]. They found that a discriminatory-price auction is more suitable in terms of maximizing revenue for the government, but an uniform-price auction is better in terms of fairness to bidders, especially small bidders. The crisis of California electricity market in 2000 reminds us that a new market mechanism, if it is not fully tested before practice, can often produce the unexpected impact on the entire economy. Because CET is not fully implemented in China up to now [22], in order to analyze its potential impact on China power sector under an experiment framework which can be repeated, this paper constructed an agent-based model which is named as CETICEM (CET Introduced China Electricity Market) and consider electricity price elasticity and carbon market clearing in China, which are very important in intra-state CET. We also compare emission-based allocation and output-based allocation, which is of particular significance for domestic carbon market mechanism design. 3. Methodology: CETICEM 3.1. Model settings In the model, six types of agents are modeled: a power grid company, power producers, other industries, consumers, the government and a carbon market maker. And two markets, the electricity and carbon markets are also modeled (as shown in Fig. 1).
R.-G. Cong, Y.-M. Wei / Energy 35 (2010) 3921e3931
3923
Definitions of the parameters and variables in equation (1) are shown in Tables 1 and 2. Expected profitability is based on the ratio of discounted net cash flows in each period and total investment costs. Different from previous studies, we consider the impact of carbon quota (ceri,j(t) cp(t)) and capital rate (infl). In equation (1), carbon quota rights can be seen as revenue because it is free allocated by government. At the same time, producers may prefer environmentally friendly electricity generation technology (for which mj is smaller). The construction of nuclear plants does not only depend on producers’ attitudes, but also on national policy. The construction of wind plants and hydropower stations depends on available land and water resources, respectively. The plants are graded, the ones with higher grade having a greater probability of being invested in. We assume that there are three types of producers: profithunting producers, environmentally friendly producers and neutral producers, with weights of profit-hunting and environmental friendliness as follows: (0.7, 0), (0.3, 0.4) and (0.5, 0.2), respectively. For example, the weight on the expected profitability for profithunting producers is 0.7, while it is 0.3 for environmentally friendly
3.2. The action rules setting 3.2.1. Power producers Power producers aim to maximize their profits. The main actions of the power producers can be seen in Fig. 2. At the beginning of each period, power producers obtain carbon quotas and determine their electricity supplies. Based on the carbon needed for electricity generation, producers decide to buy or sell carbon quotas. At the end of each period, based on the supply and demand of the market and of the operations themselves, the producers make decisions to establish or close power plants. Which type of plant will be invested in is based on multicriteria. The size of plant will be decided according to the criteria for meeting the electricity demand. The criteria include economic, environmental friendliness, nuclear fear, and the constraints of land and water resources. The economic criterion indicates the expected profitability of a power plant. The expected profitability of producer i’s plant j, epi,j, can be calculated in equation (1):
ceri;j ðtÞ cpðtÞ þ capaj pðtÞ capaj nj fpj ðtÞ capaj mj cpðtÞ ocj 1, ltj 1 A incoj epi;j ðtÞ ¼ @ 1 ð 1 þ infl þ subj incoj 1 1 1 þ infl 0
Fuel market
Other industries Carbon quotas Set quotas
Carbon market
Fuel
Cost
Price
Current installation: information Coal-fired power plants Carbon quotas
Supervise
Supply
Hydropower station…
Subsidy/Tax Government
Power producers
Set quotas
Supply
Adjust
Electricity market
Available installation: Invest
Supply Demand Price Information
Fuel price information
Price information
Supervise
Foreign power producers
Wind power plant
decision Demand information Price information
Coal-fired power plants Gas-fired power plants Solar power plant Wind power plant Hydropower Station
Power grid companies
Nuclear Power Plant Supply
Demand Consumers
Scenario generator Physical flow
Information flow Fig. 1. System structure.
Exogenous agents
(1)
3924
R.-G. Cong, Y.-M. Wei / Energy 35 (2010) 3921e3931
Obtain carbon quotas
Determine the electricity supply which is shown in the section “electricity market”
Buy fuels
Calculate carbon needed for generation
No
Carbon needed
Buy quotas
>Available quota
Yes
Is there a shortage of
No
electricity supply?
Sell quotas
Yes Multi-criterion decision
No
Is the plant at end of its life
or
making
a
continuous loss?
Yes Close the plant
Fig. 2. Flow chart of power producers’ actions.
producers. The weight on the expected profitability for neutral producers is 0.5, which means their relatively neutral attitudes to the economic criterion. In reality, profit-hunting producers may refer to private power companies whose objective is only profit maximization. Environmentally friendly producers may refer to public power companies. Not only profit maximization but also environmental protection is their objectives. Neutral producers fall in between two types of producers above. While investing, producers will firstly decide their type of producer according to the probability vector Pi ¼ {pp,pe,pn}, where pp,pe,pn represent the probabilities of turning to profit-hunting producers, environmentally friendly producers and neutral producers. They will then adjust their vectors at the end of each period based on the rules described in Table 3. For example, when
producer’s type is profit-hunting, if it can make profits, the probability of turning to profit-hunting is increased by D, the other two probabilities are decreased by D/2. The rules of probability adjusting are inspired by reinforcementlearning theory where agents take actions to maximize their longterm utilities. So the action which can get larger utility is more likely adopted in the future. Through the reinforcement-learning mechanism mentioned above, we can study whether different market design will shift consumers’ preferences. 3.2.2. Government The government’s main aim is to reduce emissions and stabilize the electricity price. Its actions include subsidizing environmentally friendly technology, laying the carbon allocation plan, setting the
R.-G. Cong, Y.-M. Wei / Energy 35 (2010) 3921e3931 Table 1 Description of parameters.
Table 3 The rules of probability adjusting.
Parameter Description
Parameter Description
a
b
g q capaj Infl mj
nj subj
GDP elasticity of electricity demand The decrease in the rate of the quota Proportional coefficient Capacity of electricity plant j Capital interest rate Carbon needed for plant j’s unit of electricity Fuel needed for plant j’s unit of electricity Subsidy for plant j, which is environmentally friendly
l ave_mar incoj ltj mari
ocj
Price elasticity of electricity demand Proportional coefficient The average profit rate of the electricity industry Investment cost for plant j Plant j’s life The ratio of producer i’s profit rate and the electricity industry’s profit rate Operation cost for plant j in one period
Producer’s type when investing
Whether the plant can make profits?
Probability adjusting
Profit-hunting producers Environmentally friendly producers Neutral producers
Yes No Yes No Yes No
P)fpp þ D; pe D2 ; pn D2 g P)fpp D; pe þ D2 ; pn þ D2 g P)fpp D2 ; pe þ D; pn D2 g P)fpp þ D2 ; pe D; pn þ D2 g P)fpp D2 ; pe D2 ; pn þ Dg P)fpr þ D2 ; pe þ D2 ; pn Dg
su ðt 1Þ ceri ðtÞ ¼ P i to capðtÞ g sui ðt 1Þ
sui(t 1) is producer i’s electricity supply in period t 1. In this case, ceri,j(t) in (1) is defined as equation (5):
sui;j ðt 1Þ ceri;j ðtÞ ¼ P P to capðtÞ g sui;j ðt 1Þ
national nuclear policy and controlling the electricity price. Carbon allocation in our model focuses on two of these actions: (I) Emission-based allocation: The previous year’s emission proportion is the standard used for setting emission quotas in the following year, which is the current quota allocation mechanism in the EU ETS. Producer i’s emission quota, ceri(t), can be calculated in equation (2):
emi ðt 1Þ ceri ðtÞ ¼ P i to capðtÞ g emii ðt 1Þ
(5)
j
And the relationship between ceri(t) and ceri,j(t) is shown in equation (4):
ceri ðtÞ ¼
X
ceri;j ðtÞ
j
Although in the first phase of EU ETS, other allocation modes, such as auction, have been in practice. But there were only three countries (Ireland, Hungary and Lithuania) choosing auction. And the percentages of carbon allowances auctioned are relatively few. Therefore, we choose two free grandfathering allocations which drew the most attentions and try to compare them.
(2)
i
where emii(t 1) is producer i’s carbon emissions in period t 1; to_cap(t) is total carbon quota in period t; and g is the decrease in the rate of the quota. ceri,j(t) in (1) is defined as equation (3):
mj sui;j ðt 1Þ ceri;j ðtÞ ¼ P P to capðtÞ g mj sui;j ðt 1Þ
(4)
i
i
i
3925
(3)
j
where sui,j(t 1) is electricity supply of producer i’s plant j in period t 1. (II) Output-based allocation: The supply proportion of the previous year’s electricity is the standard used for setting emission quotas for the following year, as shown in equation (4):
3.2.3. Industry The industry’s aim is to sell products to maximize profit. We use an agent to represent industry and assume its demand is determined by an exogenous scenario. Because power sector is our focus, industry agent’s carbon demand is set exogenously here. We will expand our object of study in the future. 3.2.4. Power grid company The action of the power grid company is setting the transmission and distribution tariff. In 2007, the average transmission and distribution tariff was 160.12 yuan/KKWH, which accounted for 31.49% of the electricity price [23]. Therefore, in the model retail electricity price equals to 131.49% of average electricity price. 3.2.5. Consumers According to different sale prices, we divide consumers into two categories: small and large electricity consumers. Small consumers need to bear the electricity price, and large consumers need to
Table 2 Description of variables. Variables
Description
Variables
Description
ave_ep(t) bid_ave(t) car_s(t) ceri,j(t) D(t) epi,j(t) GDP(t) pool(t) sui,j(t) total_cap(t) to_cap(t)
The average electricity price The average market price Total carbon supply in period t Carbon quota of producer i’s plant j Electricity demand in period t The electricity price for producer i’s plant j GDP in period t The cumulative imbalance in period t Electricity supply of producer i’s plant j in period t Total available quotas in period t Total carbon quota
bidi,j(t) car_d(t) ceri(t) cp(t) emii(t) fpj(t) P(t) si,j(t) sui(t) total_emission(t) w_sij(t)
The bid price of plant j owned by producer i in period t Total carbon demand in period t Producer i’s emission quota Carbon price in period t Producer i’s carbon emissions in period t Fuel price for plant j Retail electricity price in period t The actual supply of plant j owned by producer i in period t Producer i’s electricity supply in period t Emissions in period t The electricity supply of producer i’s plant j in period t
3926
R.-G. Cong, Y.-M. Wei / Energy 35 (2010) 3921e3931
Fig. 3. Electricity prices corresponding to different mechanisms (GDP growth, 3%; interest rate, 3%; reduction of quota, 2%). Note: we define GDP growth is 3%, because it will show the convergence trend in the long term. We also do the sensitivity analysis for different GDP and get the similar results.
cover the electricity price and capacity price. We use two agents to represent consumers. China’s electricity demand is mainly driven by gross domestic product (GDP). Meanwhile, as a commodity, its demand is certainly affected by its price. Therefore, we assume electricity demand is decided by GDP growth and its price changes accordingly. The consumers’ decisions are shown in equation (6):
3.2.6. Carbon market maker The primary function of the carbon market maker is to provide market liquidity and balance the supply and demand of the carbon market. The cumulative imbalance in period t, pool(t), can be calculated as in equation (7):
GDPðtÞ D t ¼ Dðt 1Þ 1 þ 1 a 1 GDPðt 1Þ PðtÞ 1 b þ Pðt 1Þ
where car_d(t) refers to total carbon demand in period t; car_s(t) is total carbon supply in period t; and pool(t) > 0 represents cumulative CO2 demand is larger than CO2 supply, and vice versa.
(6)
where GDP(t) refers to GDP in period t; P(t) represents retail electricity price in period t; a is the GDP elasticity of electricity demand; and b is the retail price elasticity of electricity demand.
Table 4 Comparison of electricity prices under different mechanisms. Electricity price
CET introduced
No CET introduced
Ratio
Electricity price at the end of 100 periods (yuan/KKWH) Maximum electricity price (yuan/KKWH) Minimum electricity price (yuan/KKWH) Average electricity price (yuan/KKWH) Standard deviation (yuan/KKWH) Standard deviation coefficient
1162.6
947.26
1.23
1170.38 486 876.44 176.10 0.20
965.61 486 780.72 150.13 0.19
1.12 1.3 1.04
poolðtÞ ¼ poolðt 1Þ þ car dðtÞ car ðtÞ
(7)
3.3. The market rules 3.3.1. Electricity market A two-part electricity price system operates in China, comprising the electricity price and capacity price. The capacity price is based on the average investment cost of generating units, which aims to compensate for investment cost and finance charges. The electricity price is determined by the market. This paper sets the rules as follows: in period t, firstly, each plant decides the electricity supply, w_si,j(t). If the expected return of generation is less than selling the carbon quota, the electricity supply is zero; otherwise, the supply equals its capacity. The average market price, bid_ave(t), can be calculated in equation (8):
Pn bid aveðtÞ ¼
i¼1
Pm
Pn
w si;j ðtÞ bidi;j ðtÞ Pm j ¼ 1 w si;j ðtÞ
j¼1
i¼1
Fig. 4. Electricity price at the end of 100 periods corresponding to different GDP growth rates.
(8)
R.-G. Cong, Y.-M. Wei / Energy 35 (2010) 3921e3931
3927
Fig. 5. Power source structure when CET is introduced.
As shown in equation (8), the average market price is weighed to the capacity of the bids. The bid price of plant j owned by producer i is composed of cost and profit, shown in equation (9):
bidi;j ðtÞ ¼ fpj ðtÞ capaj ðtÞ þ ocj ðtÞ þ capaj ðtÞ mj cpðtÞ ð1 þ mari ave marÞ=capaj ðtÞ (9) where ave_mar denotes the average profit rate of the electricity industry, which is set exogenously; and mari is the ratio of producer i’s profit rate and the electricity industry’s profit rate. Based on the argument that the bidding price determines the amount of capacity sold, the actual supply of plant j owned by producer i is shown in equation (10):
si;j ðtÞ ¼ w si;j ðtÞ Pn
i¼1
DðtÞ Pm j¼1
w si;j ðtÞ
bid avei ðtÞ bidi;j ðtÞ
(10)
In period t, the average electricity price, ave_ep(t), is shown in equation (11):
ave epðtÞ ¼ bid aveðtÞ e
Pn
q to DðtÞ
i¼1
Pm j¼1
w si;j ðtÞ
where q is a proportional coefficient, which reflects price fluctuations when there is an imbalance between supply and demand. As mentioned above, the bids of power plants determine the frequency of their usage. In reality, the actual price received for the electricity produced by a specific power plant is related to the height of its bid price. The electricity price for each plant is shown in equation (12):
epi;j ðtÞ ¼ ave epðtÞ
bidi;j ðtÞ bid aveðtÞ
Some validations of equations above can be found in Chappin’s research [24]. 3.3.2. Carbon market We assume that the carbon price is affected by carbon supply and demand. The current carbon price is affected by the total available quotas and emissions of the last period (current emission is not available). In period t, carbon price, cp(t), is as shown in equation (13):
cpðtÞ ¼ cpð0Þ eðlðtotal (11)
(12)
emissionðt1Þtotal capðtÞþpoolðtÞÞÞ
ð1 þ et Þ; et wN 0; 0:012
Gas
Coal
Wind
Solar Nuclear Hydro
Fig. 6. Power source structure when CET is not introduced.
(13)
3928
R.-G. Cong, Y.-M. Wei / Energy 35 (2010) 3921e3931
Table 5 Comparison of power source structures under different schemes. CET introduced The beginning period of large-scale gas 2 power generation (Proportion > 8%) Final portfolio (%) Hydropower 6.83 Solar power 24.27 Wind power 0.31 Solar power 25.25 Coal-fired 17.79 power Gas-fired power 25.53 The absolute amount Hydropower 441477.8 (MW) Wind power 20225
No CET introduced 34 11.13 16.51 0.50 11.10 36.13 24.64 453848.2 20205.9
where l is a proportional coefficient. In addition, a random disturbance, et, is also added to the model to explain other random factors and cp(0) is the initial value for the carbon price, which is set to 100 according to the EU ETS and exchange rate. 3.4. Model implementation The software used here is NetLogo which is a cross-platform multiagent programmable modeling environment. We created 90 agents as electricity power producers, each producer having a number of plants. Relevant parameters’ value of power plants are shown in the appendix Table 1. The agents evolve over time by action and interaction. The total simulation length is 100 periods. The model procedure is repeated until the end of the simulated period is reached.
grid price for wind power is 617.58 yuan/KKWH; for nuclear power is 436.23 yuan/KKWH; for thermal power is 346.33 yuan/KKWH; for hydropower is 244.04 yuan/KKWH. We can see from Fig. 3 that compared to no CET, introducing CET has the following effects: increase in the level and fluctuation of the electricity price in most cases; at the end of 100 periods (here one period means one year), the electricity price is 1.23 times higher; and fluctuation of the electricity price is 1.04 times higher (see Table 4). Therefore, introducing CET internalizes costs and transfers the volatility of the carbon price to the electricity market. Next, we study the impact of CET on the electricity price under different GDP growth rates. We do a simulation of the electricity price at the end of 100 periods corresponding to different GDP growth rates. The results are shown in Fig. 4. We can see from Fig. 4 that if no CET is introduced, when the GDP growth rate is low, the electricity price growth rate is high; and when the GDP growth rate is high, the electricity price growth rate shows a decreasing trend. On one hand, economic development encourages electricity consumption and drives up the electricity price. On the other hand, increase in the electricity price suppresses consumption. Therefore, economic development has both positive and negative impact on electricity consumption, which in return reflects the trend of decrease in electricity price growth. However, electricity price growth shows an increasing trend when CET is introduced. This is because, in addition to the above impact, economic development also demands carbon, which pushes up the carbon price. The carbon price increase is reflected in the electricity price, pushing up the electricity price. Therefore, to address China’s growing economy, the Chinese government should establish regulatory measures if CET will be introduced.
4. Experiment design and analysis 4.2. Impact of CET on power source structure The main assumptions underlying the model are an electricity market with perfect competition, a static number of electricity producers and a limited available power plant types. Main input data are listed in appendix Table 2. 4.1. Impact of CET on electricity price First of all, we explore the impact of introducing CET on the electricity price in China. The emission-based allocation, which was used in EU ETS, is chose here as the baseline scenario. We obtain different electricity prices corresponding to whether CET is introduced or not. With due consideration of availability of data, 2007 is chose as the base year. Data from China state electricity regulatory commission are used to calibrate the model. In 2007, average on-
Next, we consider the impact of CET on China’s power source structure. The emission-based allocation is also chose here. We can see from Figs. 5 and 6 that introducing CET brings forward large-scale natural gas power generation. At the same time, we can see from the final power source structure (Table 5) that there is an increase in the proportion of environmentally friendly power generation technology (nuclear power, solar power and gas-fired power). Among them, solar power, which provides environmental protection but is expensive, develops a lot after introducing CET. Due to land and water constraints, the proportions of wind power and hydropower, respectively, do not increase after CET is introduced, although, the absolute amount of wind power increases slightly.
Fig. 7. Electricity price and carbon price of two carbon quota allocations.
R.-G. Cong, Y.-M. Wei / Energy 35 (2010) 3921e3931
3929
Table 6 Group statistics of economic producers. Type of producer
Mean
Std. deviation
Minimum
Maximum
Emission-based Output-based
0.394 0.328
0.053 0.0534
0.3 0.2
0.6 0.5
Table 7 Independent sample t-test for proportion of economic producers.
Projected proportion of profit-hunting Producers
Equal variances assumed Equal variances not assumed
Levene’s test for equality of variances
t-test for equality of means
F
df
Sig.
t
0.369 0.544 8.845 198
Sig. Mean (2difference tailed)
Fig. 8. Impact of introducing CET on final producers’ proportions.
0.000 0.0666
8.845 197.989 0.000 0.0666
Table 8 Group statistics of environmentally friendly producers. Type of producer
Mean
Std. deviation
Minimum
Maximum
Emission-based Output-based
0.266 0.327
0.048 0.051
0.14 0.21
0.44 0.48
4.3. Comparison of emission-based allocation and output-based allocation As can be seen from Fig. 7, the evolution of the electricity price and carbon price of the two allocation mechanisms is similar. However, the electricity price of output-based allocation is lower because of the lower carbon price. We would like to know what the micro mechanism for this macro phenomenon is and the difference of producers under both market mechanisms. We analyze the final producers’ proportions according to the type of allocation, i.e., emission-based and output-based allocations (initial proportions: economic producers, 0.3; environmentally friendly producers, 0.3; neutral producers, 0.4). The results are as follows: As can be seen from Table 6, when emission-based allocation is introduced, the proportion of economic producers (0.394) is larger than the proportion when output-based allocation is introduced (0.328). As seen from Table 7, the probability of F is 0.544 for projected proportion of profit-hunting producers, which is insignificant at the 5% level. Thus, we believe that equal variances assumed under two allocations are not rejected. Next, let us look at the results corresponding to assumptions of equal variances. The value of t is 8.845, with probability is 0.000. The result rejects the null hypothesis and shows that there is a significant difference between the proportions of economic producers under the two allocation mechanisms. Table 9 Group statistics of neutral producers. Type of producer
Mean
Std. deviation
Minimum
Maximum
Emission-based Output-based
0.339 0.345
0.051 0.0538
0.19 0.22
0.456 0.49
We do a similar analysis on neutral and environmentally friendly producers (intermediate results are shown in Tables 8 and 9). The robust test is also done for different initial proportion settings. The final results are shown in Fig. 8. As seen in Fig. 8, when emission-based allocation is introduced, the proportion of economic producers (0.394) is higher compared to output-based allocation (0.328). In other words, emission-based allocation causes producers to turn to less environmentally friendly technologies. In essence, emission-based allocation allocates more carbon quotas to producers who emit more. This is an incentive for emissions. From a broader sense, emission-based allocation is designed for compensating bodies which are affected by emission control [25e27]. Output-based allocation encourages producers to be environment-friendly. Therefore, total emissions are lower, resulting in a lower carbon price. In this perspective, in the Chinese domestic carbon market, for the same sector, output-based allocation is more efficient for reducing emissions. The results here may be different from the general opinion which was established in standard market theory that the method of allocation of a given number of emission certificates should not affect the equilibrium price. Because the producers in our model can shift their investment preferences in the two allowance allocations, output-based allocation makes producers tend to be environment-friendly, which will decrease total carbon demand and carbon price. 5. Conclusions and future work In this paper, we studied the impact of introduction of CET on China’s power sector using an agent-based model. The main findings are as follows: (1) CET internalizes the external environmental cost, which increases the average electricity price (12%). At the same time, the volatility of the carbon market is also transferred to the electricity market (fluctuations excluding the impact of mean increases of 4%). When there is no introduction of CET, the electricity price shows a declining trend as the economy develops. However, introduction of CET significantly increases the electricity price. Therefore, to address China’s high-speed development, the Chinese government should establish specific regulatory measures after introducing CET. (2) CET influences the relative costs of different power generation technologies through the carbon price, which would have a significant impact on the power source structure.
3930
R.-G. Cong, Y.-M. Wei / Energy 35 (2010) 3921e3931
Introduction of CET would cause large-scale natural gas generation. The final proportion of environmentally friendly power generation technologies such as nuclear power and natural gas power would increase. Environmentally friendly solar power would develop significantly after CET is introduced (final proportion, 14% increase). In contrast, the final proportion of coal-fired power, whose emissions are high, would decrease significantly by 18%. (3) Emission-based allocation produces both a higher electricity price and higher carbon price, compared to output-based allocation. This is because under the latter allocation, producers would tend to be more environmentally friendly. Compared to emission-based allocation, output-based allocation would be more conducive to environmental protection. Therefore, output-based allocation should be considered in the design of the Chinese domestic CET market. As we know, this paper is the first one which studied the potential impact of CET on China power sector in a computable framework. It will provide China government and related decisionmakers a quantity tool for designing carbon market. This paper also firstly explains which free allocation is better and the micro mechanism behind it, which is a necessary supplement to existing literatures. This paper does not endogenously model the Chinese fuel market. We also do not consider entry and quit mechanisms for producers. Demand elasticity of electricity is constant. These are left to address in future research. Acknowledgements The work in this paper is part of “National carbon market mechanisms design” which aims to promote China’s reduction. The authors gratefully acknowledge the financial support from the National Natural Science Foundation of China under grant Nos.70733005, 70701032. We also would like to thank Professor Henrik Lund and the anonymous referees for their helpful suggestions and corrections on the earlier draft of our paper according to which we improved the content. Appendix Table 1. Parameters’ values of power plants.
References [1] EC. Directive 2003/87/EC Establishing a scheme for greenhouse emission allowance trading within the community and amending council directive 96/ 61/EC, Report; 2003 See also, http://europa.eu/legislation_summaries/energy/ european_energy_policy/l28012_en.htm#AMENDINGACT. [2] Svendsen GT. The idea of global CO2 trade. European Environment 1999;9:232e7. [3] Wei YM, Liu LC, Fan Y, Wu G. China energy report: CO2 emissions research. Beijing: Science Press. 2008 See also, http://www.sciencep.com/s_single.php?id¼14944. [4] NBSdNational Bureau of Statistics of China. China energy statistical yearbook 1997e2008. Beijing: China Statistical Press; 2008. [5] Bohringer C, Lange A. On the design of optimal grandfathering schemes for emission allowances. European Economic Review 2005;49:2041e55. [6] Cramton P, Kerr S. Tradeable carbon permit auctions: how and why to auction not grandfather. Energy Policy 2002;30:333e45. [7] Burtraw D, Palmer K, Bharvirkar R, Paul A. The effect of allowance allocation on the cost of carbon emission trading. RFF Discussion Paper; 2001. [8] Paul A, Burtraw D. The RFF haiku electricity market model. RFF Discussion Paper; 2002. [9] Haurie A, Viguier L. A stochastic dynamic game of carbon emissions trading. Environmental Modeling and Assessment 2003;8:239e48. [10] Carlén B. Market power in international carbon emissions trading: a laboratory test. The Energy Journal 2003;24:1e26. [11] Bosello F, Roson R. Carbon emissions trading and equity in international agreements. Environmental Modeling and Assessment 2002;7:29e37. [12] Keats K, Neuhoff K, Martinez KK. Allocation of carbon emission certificates in the power sector: how generators profit from grandfathered rights. Climate Policy 2005;5:61e78. [13] Szabó L, Hidalgo I, Ciscar JC, Soria A. CO2 emission trading within the European Union and Annex B countries: the cement industry case. Energy Policy 2006;34:72e87. [14] Chappin EJL, Dijkema GPJ. On the impact of CO2 emission-trading on power generation emissions. Technological Forecasting and Social Change 2009;76:358e70. [15] Ma T, Nakamori Y. Modeling technological change in energy systemsefrom optimization to agent-based modeling. Energy 2009;34:873e9. [16] Bunn DW, Oliveira FS. Evaluating individual market power in electricity markets via agent-based simulation. Annals of Operations Research 2003;121(1):57e77. [17] Ehlen MA, Scholand AJ, Stamber KL. The effects of residential real-time pricing contracts on transco loads, pricing, and profitability: simulations using the NABLE(TM) agent-based model. Energy Economics 2007;29(2):211e27. [18] Bunn DW, Oliveira FS. Agent-based analysis of technological diversification and specialization in electricity markets. European Journal of Operational Research 2007;181(3):1265e78. [19] Hamalainen RP, Mantysaari J, Ruusunen J, Pierre-Olivier P. Cooperative consumers in a deregulated electricity market e dynamic consumption strategies and price coordination. Energy 2000;25(9):857e75. [20] Bernal-Agustin JL, Contreras J, Martin-Flores R, Conejo AJ. Realistic electricity market simulator for energy and economic studies. Electric Power Systems Research 2007;77:46e54. [21] Cong RG, Wei YM. Auction design for the allocation of carbon emission allowances: uniform or discriminatory price? International Journal of Energy and Environment 2010;1(3):533e46. [22] Raufer R, Li S. Emissions trading in China: a conceptual ‘leapfrog’ approach? Energy 2009;34(7):904e12.
Type
Life (ltj)
Installed capacity (capaj) (MW)
Unit cost of power generation (fpj(t)nj) (yuan/KKWH)
Construction cost (incoj) (yuan/KKWH)
Carbon emission (mj) (ton CO2/KKWH)
Hydrostation Nuclear plant Wind plant Solar plant Coal-fired plant Gas-fired plant
30 40 20 20 20 20
600 300 10 20 300 120
240 440 620 1900 350 800
6000 12,000 8000 10,000 4000 4000
0 0 0 0 1.3 0.7
Source: [28,29]. Some data are from the historical data of plants that have been established. Others are authors’ calculations according to actual survey data from Chinese power plants.
Table 2. Main input data. Parameter
Value
Parameter
Value
a l q
0.8 0.005 0.001 0.03 0 100
b g
0.1 0.95 0.1 1 0
Infl ocj cp(0)
ave_mar mari subj
R.-G. Cong, Y.-M. Wei / Energy 35 (2010) 3921e3931 [23] SERCdState Electricity Regulatory Commission. Monitoring report on the implementation of tariff 2008. Internal Report; 2008. See also, http://www. serc.gov.cn/zwgk/jggg/200809/W020080912334874610579.doc. [24] Chappin EJL. Carbon dioxide emission trade impact on power generation portfolio, agent-based modelling to elucidate influences of emission trading on investments in Dutch electricity generation. Master Thesis, Delft University of Technology, Delft; 2006. [25] Tietenberg T. The tradable permits approach to protecting the commons: lessons for climate change. Oxford Review of Economic Policy 2003;19:400e19. [26] Harrison D, Radov D. Evaluation of alternative initial allocation mechanisms in a European Union greenhouse gas emissions allowance trading scheme. Prepared for DG Environment, European Commission; 2002. [27] Ahman M, Burtraw D, Kruger J, Zetterberg L. A ten year rule to guide the allocation of EU emission allowances. Energy Policy 2006;35:1718e30.
3931
[28] Peng C. China hydropower project in the 21st century. Beijing: China Water Power Press; 2006. [29] SERCdState Electricity Regulatory Commission. Electricity regulatory annual report. Internal Report; 2007 See also, http://www.serc.gov.cn/zwgk/jggg/ 200808/t20080801_9841.htm.
Mr. Rong-Gang Cong is a Ph.D. candidate in Management Science at the Institute of Policy and Management, Chinese Academy of Sciences, China.
Dr. Yi-Ming Wei is a Professor at the Center for Energy and Environmental Policy Research, Beijing Institute of Technology (BIT), and is the Dean of School of Management and Economics, BIT. He was a visiting scholar at Harvard University, USA.